How does the bias point determine amplifier class?

The amplifier class is defined by the fraction of the RF cycle during which the transistor conducts current, which is set by the Q-point position relative to the device pinchoff voltage. Class A: IDQ = IDSS/2 (50% of saturated drain current). The device conducts for the entire 360 degrees of the RF cycle. The Q-point is at the midpoint of the IV curve, allowing symmetric voltage and current swings. Maximum linearity (no clipping at small signals), but the device dissipates DC power even with no RF input. Drain efficiency: eta = Pout/(PDC) = (Vswing/2)^2 / (2 * RL * VDD * IDQ). Maximum theoretical efficiency: 50% when Vswing = VDD - Vknee and the current swings from 0 to 2*IDQ. Practical: 25-35% due to knee voltage, non-ideal current waveform, and backoff for linearity. Class AB: IDQ = 0.05-0.3 IDSS. The device is biased just above pinchoff, conducting for 180-360 degrees depending on the exact IDQ. Below the linear region of Class A but with improved efficiency (35-60%). The conduction angle decreases as IDQ decreases, trading linearity for efficiency. Most wireless infrastructure PAs use deep Class AB (IDQ approximately 5-10% IDSS) with digital predistortion (DPD) to compensate for the reduced linearity. Class B: IDQ approximately 0 (biased at pinchoff). Conducts for exactly 180 degrees. The output is a half-sinusoid that requires a resonant tank or complementary push-pull topology to reconstruct the full waveform. Theoretical efficiency: 78.5% (pi/4). Practical: 50-65%. Generates significant even-order harmonics (2nd, 4th). Class C: IDQ = 0 (biased below pinchoff). Conducts for less than 180 degrees. Higher efficiency (up to 85-90% theoretical) but highly nonlinear. Used only for constant-envelope modulations (FM, CW) or as frequency multipliers.

What is load line analysis and why does it matter?

Load line analysis is a graphical technique that determines the maximum RF output power by plotting the transistor IV curves and overlaying the impedance presented to the device by the matching network. The DC load line has slope -1/RDC (where RDC is the total DC resistance in the drain/collector path) and passes through the Q-point and the supply voltage VDD on the voltage axis. The RF load line has slope -1/RL (where RL is the load resistance presented to the transistor at the fundamental frequency) and also passes through the Q-point. For Class A: maximum RF power is delivered when the voltage swings symmetrically from Vknee to 2*VDD - Vknee and the current swings from 0 to 2*IDQ. This gives: RL_opt = (VDD - Vknee) / IDQ, and Pout_max = (VDD - Vknee)^2 / (2 * RL_opt) = IDQ * (VDD - Vknee) / 2. Example: GaN HEMT at VDD = 28V, Vknee = 4V, IDQ = 500 mA. RL_opt = (28-4)/0.5 = 48 ohms. Pout_max = 0.5 * 24 / 2 = 6W (37.8 dBm). The matching network must transform the system impedance (typically 50 ohms) to RL_opt at the device reference plane. For Class B/AB: the load line is not symmetric. The voltage swing extends from Vknee to 2*VDD - Vknee (same as Class A), but the current is clipped at zero for part of the cycle. The optimal load resistance is higher: RL_opt(B) = 2 * (VDD - Vknee) / Imax, where Imax is the peak current at the top of the current waveform.

How does bias point affect noise figure?

For low-noise amplifiers (LNAs), the bias point directly determines the noise figure through its effect on the transistor noise parameters. The minimum noise figure NFmin of a FET depends on drain current ID approximately as: NFmin approximately equals 1 + (2*f/fT) * sqrt(gds0/gm), where f is the operating frequency, fT is the unity-gain cutoff frequency, gds0 is the output conductance at VDS = 0, and gm is the transconductance. gm increases with ID up to a saturation point (typically 30-50% of IDSS), while fT also increases with ID up to a peak. The optimal noise bias point (ID for minimum NFmin) is typically 10-20% of IDSS, which is lower than the optimal gain bias point (30-50% IDSS). At the noise-optimal bias: NFmin may be 0.3-0.5 dB lower than at the gain-optimal bias, but the associated gain is 2-5 dB lower. The system designer must balance NF versus gain based on the cascade noise analysis (Friis formula). For a first-stage LNA where NF dominates system sensitivity, biasing at 10-15% IDSS is preferred. For a second-stage amplifier where gain reduces the noise contribution of subsequent stages, biasing at 30-50% IDSS maximizes overall system SNR. Temperature also affects NFmin: drain current at the noise-optimal bias shifts with Vp drift (-2 mV/C), so the LNA bias circuit must maintain the noise-optimal ID over temperature, not a fixed VGS. Active bias circuits that servo ID to a fixed value (via a sense resistor and op-amp) provide the best NF stability over temperature.

Amplifier Design / Active

Bias Point

/BYE-us poynt/ — Q-point

The DC operating point (I_DQ, V_DSQ) of an RF transistor on the IV characteristic curves. Determines amplifier class (A/AB/B/C), linearity, efficiency, noise figure, and maximum output power. Load line analysis: R_L,opt = (V_DD − V_knee)/I_DQ. P_out,max = I_DQ(V_DD − V_knee)/2 for Class A. Noise-optimal bias: 10–20% I_DSS (0.3–0.5 dB better NF than gain-optimal).

Class A: I_DQ = I_DSS/2

η_max: 50% (Class A)

NF optimal: 10–20% I_DSS

Understanding the Bias Point

The bias point defines the transistor's quiescent state before any RF signal is applied. Its position on the IV curves directly determines how the device will process the RF waveform: the conduction angle (fraction of the RF cycle with current flow), the available voltage swing (limited by the knee voltage and supply rail), and the noise performance (which is minimized at a lower current than the gain-optimal point).

Load line analysis translates the Q-point and load impedance into maximum RF output power. The RF load line, plotted through the Q-point with slope −1/R_L, defines the voltage and current excursion during the RF cycle. Optimal load impedance maximizes the area enclosed by the load line within the safe operating region of the IV curves.

Load Line and Power

Optimal Load (Class A):
R_L,opt = (V_DD − V_knee) / I_DQ

Maximum Output Power:
P_out,max = I_DQ(V_DD − V_knee) / 2
P_DC = V_DD × I_DQ
η_max = (V_DD − V_knee) / (2V_DD) ≤ 50%

Example (GaN HEMT):
V_DD = 28 V, V_knee = 4 V, I_DQ = 500 mA
R_L,opt = 24/0.5 = 48 Ω
P_out,max = 0.5 × 24/2 = 6 W (37.8 dBm)

Q-Point by Amplifier Class

Class	I_DQ / I_DSS	Conduction	η_max	Linearity	Application
A	50%	360°	50%	Best	LNA, driver
AB	5–30%	200–350°	35–60%	Good (with DPD)	Base station PA
B	~0%	180°	78.5%	Moderate	Push-pull
C	0% (below V_p)	<180°	85–90%	Poor	FM, CW, multiplier

Noise vs. Gain Optimization

Bias Target	I_DQ / I_DSS	NF_min	Gain	Use Case
Noise-optimal	10–20%	Minimum	−2 to −5 dB	1st-stage LNA
Gain-optimal	30–50%	+0.3–0.5 dB	Maximum	2nd stage, driver
Power-optimal	50%	+1–3 dB	Max P_out	PA

Common Questions

Frequently Asked Questions

How does Q-point set amplifier class?

Class A: I_DQ = I_DSS/2 (360° conduction, max linearity, 50% η). AB: 5–30% I_DSS (200–350°, 35–60% η). B: at pinchoff (180°, 78.5% η, needs push-pull). C: below pinchoff (<180°, up to 90% η, nonlinear, FM/CW only).

Load line analysis?

R_L,opt = (V_DD − V_knee)/I_DQ for Class A. GaN at 28 V, 4 V knee, 500 mA: R_L = 48 Ω, P_out = 6 W (37.8 dBm). Matching network transforms 50 Ω to R_L,opt. Class B: R_L = 2(V_DD − V_knee)/I_max.

Noise figure vs. bias point?

NF_min occurs at 10–20% I_DSS, 0.3–0.5 dB better than gain-optimal (30–50% I_DSS). First-stage LNA: bias for noise. Second stage: bias for gain (Friis formula reduces NF contribution). Active bias holding I_D constant over temperature provides best NF stability.

Related Terms

← Bias Network Bias Point Optical →

Amplifier Design

Precision RF Components

RF Essentials provides precision terminations and custom RF assemblies for amplifier load-pull characterization, Q-point optimization, and power amplifier design validation.

Request a Quote